The present invention is directed to a method for producing siliconized polysilicon contacts in integrated semiconductor structures, to the employment of the method for manufacturing a transistor, as well as to a transistor that can be obtained according to the inventive method.
In the manufacture of integrated semiconductor structures, the problem frequently arises that the resistance of polysilicon structures employed as interconnects must be selectively lowered. At the same time, the resistance of other polysilicon structures located in the integrated semiconductor, which is to be manufactured, should remain high since these, for example, are to be employed as resistors within the integrated circuit. In order to reduce the resistance of a polysilicon structure, the structure is often provided with a silicide layer. However, the manufacturing problem of siliconizing only specific polysilicon structures but not siliconizing others, for example those that are to be employed for resistors will occur.
The problem of selective siliconization and solutions known from the prior art shall be discussed below on the basis of a specific example of the employment in an integrated transistor. It is self-evident, however, that the inventive method can also be applied to other fields of employment in semiconductor production and is not intended to be limited to the employment in the manufacture of integrated transistors.
FIG. 1 shows a bipolar transistor in an integrated semiconductor known in the prior art. The actual, active transistor 1 is thereby composed of three juxtaposed, differently doped regions of semiconductors, the emitter region 2, the base region 3 as well as the collector region 4. Dependent on the doping, a distinction is made between pnp-transistors, and npn-transistors, whereby the sequence of the letters identifies the sequence of the doping in the emitter, base and collector region.
Over and above this, a transistor arranged in an integrated circuit also has further auxiliary structures surrounding it, and these structures serve, on the one hand, for isolating potentials and, on the other hand, for the elimination of the currents from the active transistor region 1. The emitter region 2 is connected via an emitter contact 5, which is usually composed of polysilicon, to an emitter interconnect 6 of, for example, aluminum. The base region 3 is connected via a base contact 7 to a base interconnect 8. The collector region 4, finally, is connected via what is referred to as a buried layer 9, which is located under the other structures, and an intermediate layer 10 as well as a collector contact 11 to a collector interconnect 12. Various silicon oxide insulating layers 13 as well as spacer insulators 14 serve the purpose of the electrical separation of the various, electrically conductive structures.
What is referred to as the extrinsic base resistance, which is the resistance between the base 3 and the base interconnect 8, is, in addition to the transit frequency and the base-collector capacitance, the critical transistor parameter in bipolar transistors that defines important characteristic quantities of the transistor such as its maximum oscillation frequency, its gain, its minimum noise after, its gate delay times etc. Thus, valid here, for example, are                               f          max                ≈                                            f              T                                      8              ⁢                              π                ·                                  R                  B                                ·                                  C                  BC                                                                                        (        1        )            
whereby
fmax: maximum oscillation frequency
fT: transit frequency
RB: base resistance
CBC: base-collector capacitance
or       F    min    ≈      1    +          1      β        +                  f                  f          T                    ⁢                                                                  2                ·                                  I                  C                                                            V                T                                      ⁢                                          R                B                            ⁡                              (                                  1                  +                                                            f                      T                      2                                                              β                      ·                                              f                        2                                                                                            )                                              +                                    f              T              2                                      β              ·                              f                2                                                        
whereby
Fmin: minimum noise factor
xcex2: gain
f: frequency
fT: transit frequency
IC: collector current
VT: thermal voltage
RB: base resistor apply.
In self-aligned silicon bipolar transistors, the base resistance is essentially composed of three parts, these being referred to below as RB,l, RB,e and RB,T. The inner part RB,l occurs due to the resistance of the base zone in the active transistor 1 under the emitter region 2. The external part RB,e describes the resistance of the polysilicon track 7 that forms the base content. RB,T represents the base resistance that occurs due to a lightly doped zone under the self-aligned emitter base insulation of the spacer insulator 14 at the active transistor. This region is generally referred to as link region in the literature.
Optimizations can be undertaken at all three regions in order to reduce the base resistance. For typical applications, for example, in micro-transistors or ultra high frequency circuits, the transistor is configured in an arrangement with two base contacts instead of one base contact 7 (not shown).
The second base contact can, for example, be arranged between the emitter contact and the collector contact. The reduction of the base resistance that is thus achieved, however, is purchased at the expense of an increased space requirement, higher capacitances, a higher power consumption and lower transit frequency.
The advantages of the low base resistance of a transistor having two base contacts and the small structure of a transistor having one base contact can be united when the polysilicon that serves the purpose of base contacting is siliconized, i.e. provided with a silicide layer. The significantly lower film contact of the silicide compared to the polysilicon leads to the fact that the transistor side facing away from the base contact is also connected low-impedance to the base contact via the silicide and, thus, a similar, low base resistance as given a transistor with two base contacts results. Moreover, the resistance inherent in the base contact is reduced.
In addition to being employed for the above-described function, silicide layers can also be utilized as additional wiring level. This enables the optimization of the wiring layout and, thus, of the circuit performance capability.
In a simple, previously known method for siliconization, the siliconization step is introduced immediately after the coating and structuring of the polysilicon. This method, however, has the serious disadvantage of siliconizing not only the desired base terminal regions as well as, potentially, additional interconnects but of siliconizing all uncovered polysilicon regions, i.e. even those wherein a siliconization is unwanted. However, it is precisely the silicon layers that are employed for the base contact as well as the emitter and collector terminal in integrated circuits that are also employed for realizing ohmic load resistors in the circuits. Due to the complete siliconization, the film resistance of the polysilicon layer or layers become so low that resistors (that typically exhibit values from 50 through 1000 ohms) can no longer be meaningfully produced with the assistance of these layers. When a silicide is therefore to be integrated into an integrated circuit, methods are thus necessary that prevent a siliconization in the region of the load resistors. Such methods are referred to in the literature as xe2x80x9csilicide blockingxe2x80x9d. A known method for silicide blocking is implemented with the assistance of photolithography. With the assistance of a resist mask, the dielectric that surrounds the resistor and transistor regions is covered in those regions wherein no silicide formation should occur. With an etching, the dielectric is opened in regions to be siliconized, i.e. removed, and a siliconization is subsequently implemented. The introduction of this additional photolithographic step means a clear increase in the process complexity compared to methods that do not employ a siliconization and thus leads to considerably increased process costs that often make the use of a silicide seem not meaningful despite the improvement of the transistor performance capability.
The present invention is therefore based on the object of offering a method wherein a selective, i.e. designational siliconization of polysilicon structures is possible without noticeably increased process complexity. This object is inventively achieved by a method for producing siliconized polysilicon contacts in integrated semiconductor structures, for producing a transistor and to a transistor which is produced according to the improved method.
Further advantageous aspects and developments of the invention derive from the patent claims, from the specification and from the attached drawings.
The present invention is based on the basic principle of already covering polysilicon structures with a dielectric before their structuring, this dielectric, in a following step, preventing a dielectric applied earlier from being etched off. A siliconization occurring subsequently takes place only at those locations where the dielectric was etched off.
Accordingly, the present invention is directed to a method for producing siliconized polysilicon regions in integrated semiconductor structures of a semiconductor blank having at least one structure formed in a first polysilicon layer and a layer of a first dielectric superimposed on the first polysilicon layer, said method having the following steps:
applying a second polysilicon layer onto the layer of the first dielectric;
applying a layer of a second dielectric onto the second polysilicon layer;
forming at least one predetermined structure in the layer of the second dielectric and of the second polysilicon layer;
removing the regions of the layer of the first dielectric that are not covered by the at least one predetermined structure, so that the at least one structure fashioned in the first polysilicon layer is at least partly uncovered; and
forming a silicide layer on the at least one uncovered structure.
What is thereby to be understood by a semiconductor structure or structure is thereby any interconnected region composed of a uniform material within an integrated semiconductor circuit. Dependent on previously ensuing etch-off or deposition steps, such a structure, accordingly, can also be two-dimensionally or three-dimensionally fashioned but will always have a layer-like structure which occurs in the manufacture of integrated semiconductors.
What is to be understood by a semiconductor blank is an arrangement of structures produced by a semiconductor process that are still in process of being finished, i.e. are not yet fully functional. In the present instance, what are to be particularly understood by this term are those semiconductors being fabricated wherein, following introduction of function elements, a polysilicon layer has been applied and a dielectric layer, which lies above the polysilicon layer, has been applied, so that the polysilicon layer is already structure, i.e. wherein corresponding semiconductor structures have been formed. What is to be understood by a superimposed layer in the sense of the present invention is a layer that has been applied on the semiconductor blank temporarily following another layer superimposed by this layer in the production sequence.
It is self-evident that a structure is formed not only in the first polysilicon layer or, respectively, a structure is formed in the second polysilicon layer; on the contrary a plurality or, respectively, multitude of structures will be present in the two layers, whereby the inventive method also functions when only a single structure in a polysilicon layer is to be siliconized.
The fashioning of the predetermined structure or structures in the layer of the second dielectric and of the second polysilicon layer thereby ensues simultaneously, i.e. that the predetermined structure is imaged both in the layer of the second dielectric as well as in the second polysilicon layer lying therebelow. This predetermined structure or predetermined structures can represent one or more function elements of an integrated circuit. For example, they can be resistors, interconnects or emitter contacts as well of an integrated transistor.
The predetermined structures, however, need not necessarily assume a function within the integrated circuit. On the contrary, they can also or also additionally form a covering that comes to lie over predetermined regions of the first polysilicon layer. What can be achieved in this way is that such covered regions in the first polysilicon layer are not siliconized, even when a function element of the second polysilicon layer accidentally does not come to lie above them. The formation of the at least one predetermined structure or of the predetermined structures preferably comprises the following steps:
Applying a photoresist onto the layer of the second dielectric;
Exposing the photoresist with a mask that images the at least one predetermined structure onto the photoresist;
Developing the photoresist; and
Removing regions of the layer of the second dielectric and of the second polysilicon layer that do not form the at least one predetermined structure.
What is thereby to be understood by a photoresist is a standard photoresist known in the prior art.
Since the regions of the layer of the first dielectric that are not covered must be selectively removed, the two layers of the dielectric must differ in terms of the materials employed in order to enable a selective removal of the layer of the first dielectric. Thus, the removal of the layer of the first dielectric can preferably ensue by etching with an etchant that is selective for the first dielectric. The layer of the first dielectric can, for example, contain silicon dioxide or can be composed of silicon dioxide. This can be etched off with the preferred etchant hydrogen fluoride. The layer of the second dielectric should be etchable by an etchant which is not employed as the etchant for the first layer of dielectric. For example, silicon nitride can be employed for the second dielectric, and silicon nitride will not be etched by hydrogen fluoride.
The silicide layer that is formed can be composed of silicides usually employed in semiconductor technology or, respectively, can at least contain this, for example the preferred titanium silicide, platinum silicide and/or cobalt silicide.
In a preferred embodiment, the at least one uncovered structure in the first polysilicon layer is the base contact of an integrated transistor. In order to maintain a certain resistance in the base contact, it is preferred that it is not the entire surface of the base contact that is siliconized.
The invention is also directed to a transistor that can be obtained by the above-described, inventive method.
Finally, the invention is also directed to the employment of the above-cited method for manufacturing a transistor. Preferably, such a transistor is a bipolar transistor. The method, however, can also be applied to other technologies, for example to CMOS technologies.
In addition to transistors, other siliconized structures can also be manufactured when a CMOS process is utilized.
The invention shall be explained below on the basis of an exemplary embodiment, whereby the attached drawings shall be referenced.